McNair Shah(@Mcn_S7):I think there is a general tendency to define the strength of mechinterp as ‘ability to produce cot-esque text about model reasoning’ (e.x. NLAs) I think this incorrectly ignores mechinterp research as a way to build fundamental understanding and intuition about model behavior I agree that we probably won’t get to tools that make us happy by hillclimbing the first in <1 year, but I disagree that we won’t be able to make useful leaps in our understanding per the second

McNair Shah

@Mcn_S7

AI Safety Researcher | Computer Science Undergraduate @ CMU | CEO @mcnairai

加入 June 2025

144 正在關注 532 粉絲

McNair Shah@Mcn_S7

2026.09.02 09:04

I think there is a general tendency to define the strength of mechinterp as ‘ability to produce cot-esque text about model reasoning’ (e.x. NLAs) I think this incorrectly ignores mechinterp research as a way to build fundamental understanding and intuition about model behavior I agree that we probably won’t get to tools that make us happy by hillclimbing the first in <1 year, but I disagree that we won’t be able to make useful leaps in our understanding per the second